2934
Biochemistry 1988, 27, 2934-2943
Hermes, J. D., Roeske, C. A., OLeary, M. H., & Cleland, W. W. (1982) Biochemistry 21, 5106. Hsu, R. Y. (1970) J. Biol. Chem. 245, 6675. Hsu, R. Y. (1984) in Biochemistry of Metabolic Processes (Lennon, D. L., Stratman, F. W., & Zahlten, R. N., Eds.) pp 259-272, Elsevier Science, New York. Hsu, R. Y., & Lardy, H. A. (1967) J. Biol. Chem. 242,520. Hsu, R. Y., Lardy, H. A,, & Cleland, W. W. (1967) J. Biol. Chem. 242, 5315. Jensen, W. B. (1978) Chem. Rev. 78, 1. Loo, S.,& Ennan, J. E. (1977) Biochim. Biophys. Acta 481, 279. Martell, A. E., & Smith, R. M. (1977) Critical Stability Constants, Vol. 1, Plenum, New York. Northrop, D. B. (1977) in Isotope Effects on Enzyme-Catalyzed Reactions (Cleland, W. W., O'Leary, M. H., &
Northrop, D. B., Eds.) p 122, University Park Press, Baltimore, MD. OLeary, M. H. (1980) Methods Enzymol. 64, 83. Phillips, H. O., Marcinkowsky, A. E., Sachs, S.B., & Kraus, K. A. (1977) J. Phys. Chem. 81, 679. Pocker, Y., & Janjic, N. (1987) Biochemistry 26,2597-2606. Schimerlik, M. I., & Cleland, W. W. (1977) Biochemsitry 16, 576. Schimerlik, M. I., Grimshaw, C. E., & Cleland, W. W. (1977) Biochemistry 16, 57 1. Seltzer, S.,Hamilton, G. A., & Westheimer, F. H. (1959) J. Am. Chem. SOC.81, 4018. Stokes, R. H., & Weeks, I. A. (1964) Aust. J. Chem. 17,304. Tang, C. L., & Hsu, R. Y. (1973) J. Biol. Chem. 135,237. Viega Salles, J. B., & Ochoa, S . (1950) J. Biol. Chem. 187, 849.
Isotope Effect Studies of the Chemical Mechanism of Pig Heart NADP Isocitrate Dehydrogenase? Charles B. Grissom: and W . W . Cleland* Department of Biochemistry, University of Wisconsin, Madison, Wisconsin 53706 Received August 26, 1987; Revised Manuscript Received December 18, 1987
ABSTFWT: The catalytic mechanism of porcine heart NADP isocitrate dehydrogenase has been investigated by use of the variation of deuterium and 13Ckinetic isotope effects with pH. The observed 13Cisotope effect ~ 1.OO28 at neutral pH to a limiting value of 1.040 at low pH. The limiting on V / K for isocitrate i n c r from 13Cisotope effect with deuteriated isocitrate at low pH is 1.016. This decrease in l3(V/KIC) upon deuteriation indicates a stepwise mechanism for the oxidation and decarboxylationof isacitrate. This predicts a deuterium isotope effect on V / K of 2.9, but D( V / K )at low p H only increases to a maximum of 1.08. It is not known why l3(V/KIc) with deuteriated isocitrate decreases more than predicted. The pKseen in the 13(V/KIc)pH profile for isocitrate is 4.5. This pK is displaced 1.2 pH units from the true pK of the acid/base functionality of 5.7 Seen in the pKi profile for oxalylglycine, a competitive inhibitor for isocitrate. From this displacement, catalysis is estimated to be 16 times faster than substrate dissoCiation. By use of the pHdependent partitioning ratio of the reaction intermediate oxalosuccinate between decarboxylation to 2-ketoglutarate and reduction to isocitrate, the forward commitment to catalysis for decarboxylation was determined to be 7.3 at pH 5.4 and 3.2 at pH 5.0. This gives an intrinsic "C isotope effect for decarboxylation of 1.050. 3-Huoroisocitrate is a new substrate oxidatively decarboxylated by NADP isocitrate dehydrogenase. At neutral pH, D( V/KZFmIc) = 1.45 and 13(V/K3.F-Ic)= 1.0129. At pH 5.2, '3(V/K3.F~I,) increases to 1.0186, indicating that a finite, but diminished, external commitment remains a t neutral pH. The product of oxidative decarboxylation of 3-hydroxyisacitrate by NADP isocitrate dehydrogenase is 2-hydroxy-3-ketoglutarate. This results from enzymatic protonation of the cis-enediol intermediate a t C2 rather than C3 (as seen with isocitrate and 3-fluoroisacitrate). 2-Hydroxy-3-ketoglutarate further decarboxylates in solution to 2-hydroxy-3-ketobut4;rate, which further decarboxylates to acetol. This makes 3-hydroxyisocitrate unsuitable for I3Cisotope effect studies.
p!!, heart NADP isocitrate dehydrogenase (EC l .42) catalyzes the oxidative decarboxylation of threo-D&ocitrate NADP
0006-2960/88/0427-2934$01.50/0
l
I H-C-OH
to 2-ketoglutarate and C 0 2 with reduction of NADP: tSupported by a grant from the National Institutca of Health (GM 18938). A prtliminaryreport of this work was presented at the American Society of B i o b g i d Chemists Meeting, New Orleans, LA,April 15-23, 1982 (Grissom & Cleland, 1982), and the American Chemical Society Meeting, Philadelphia, PA, September 1982 ( G r i m & Ckland, 1984). *Presentaddress: Department of Chemistry, University of California, Berkeley, CA 94720.
coo-
coo-
l. l.
+
I I H-C-H
-0OC-C-H
c=o
l
coothreo-0-
I
molr1(2+)
+
CO2 4- NADPH
2-ketoglutarate
(1)
# H-C-H
isocitrate
I H-C-H I
coo-
It is entirely specific for NADP' and the fully ionized 2R,3S 0 1988 American Chemical Society
N A D P ISOCITRATE DEHYDROGENASE ISOTOPE EFFECTS
isomer of isocitrate. The divalent metal cation most often used is Mn2+ or Mg2+,with lesser activity seen with Ni2+, Zn2+, or Cd2+. The true substrate for reaction 1 has been postulated to be the metal-isocitrate complex (Colman & Villafranca, 1972; Colman, 1972), but this may represent highly synergistic binding of metal ion and isocitrate. The reaction is reversible with C 0 2 being the oxidized carbon species used (Dalziel & Londesborough, 1968). Active enzyme centrifugation and Kurganov (1967) enzyme dilution experiments indicate the dimer is active and the monomer is inactive (Kelly & Plaut, 1981a,b). The kinetic mechanism has been thoroughly investigated by use of initial velocity, multiple inhibition, and isotope exchange studies and shown to be random sequential with catalysis more rapid than product release (Uhr et al., 1974; Northrop & Cleland, 1974). Kelly and Plaut (198 1b) showed that the kinetic mechanism changes to ordered sequential with NADP binding last at low enzyme concentrations or in the absence of NADPH or EDTA.' In addition to reaction 1, the enzyme will also catalyze the reduction or decarboxylation of oxalosuccinate (Siebert et al., 1957) as shown in eq 2 and 3. Because these activities reside coocoo-
I I -0OC-C-H
l
c=o
c=o
I
metal (2+)
I H-C-H l coo-
#
oxalosuccinate
I H-C-H I
I
H-C-H
l coo-
oxalosuccina te
(2)
coocoo-
I c=o
+ -0OC-C-H I
CO2
2-ketoglutarate
coo-
NADPH
+
H-C-H
NADPH
l I -0OC-C-H
H-C-OH
meta1(2+)
+
NADP ( 3 )
H-C-H
I
coothreo-D-
isocitrate
on the same protein, the mechanism of isocitrate oxidative decarboxylation is generally thought to be a stepwise process with hydride transfer producing enzyme-bound oxalosuccinate, which is subsequently decarboxylated. No incorporation of radioactivity into an unlabeled oxalosuccinate pool could be demonstrated when starting from isocitrate (Siebert et al., 1957). The enzyme will catalyze the exchange of the pro-S C3 methylene hydrogen of 2-ketoglutarate with solvent, indicating an enzymatic base capable of stereospecifically protonating the 2-ketoglutarate enolate resulting from decarboxylation (Rose, 1960; Lienhard, 1964). The same type of stepwise dehydrogenation, followed by metal-assisted decarboxylation, has been shown to operate with NADP and NAD malic enzymes, which are analogous oxidative decarboxylases (Hermes et al., 1982; Grissom et al., 1987). For NADP malic enzyme, an observed 13C kinetic isotope effect on V/K of 1.0336 was observed at neutral pH with Mg2+. For NADP isocitrate dehydrogenase, a l3(V/KIC) value of 0.9989 f 0.0004 was observed at neutral pH (0'Leary, 1971; Limburg & OLeary, 1977). This nearly nonexistent isotope effect is attributed to the stickiness of isocitrate I Abbreviations: Pipes, 1,4-piperazinediethanesulfonic acid; Mes, 2-(N-morpholino)ethanesulfonicacid; EDTA, ethylenediaminetetraacetic acid; Hepes, N-(2-hydroxyethyl)piperazine-N'-2-ethanesulfonic acid; OSA, oxalosuccinic acid; IC, isocitrate; IcDH, isocitrate dehydrogenase; 2-KG, 2-ketoglutarate.
VOL. 27, NO. 8, 1988
2935
(Le., catalysis is faster than the net rate constant for dissociation of isocitrate from the central enzyme-substrate complex) such that no isotopic discrimination between 12C-and 13C-containingsubstrates occurs. The deuterium isotope effect on V/KIc varies slightly with pH: D( V/KIc) is 1.0 at pH 7.45 and 1.08 at pH 4.5 (Cook & Cleland, 1981b). In this paper, we have used the variation of 13Ckinetic isotope effects with pH (Cook & Cleland, 1981a,b) and the partitioning of oxalosuccinate between reduction and decarboxylation (Grissom & Cleland, 1985) to determine the chemical mechanism and compare it to that of NADP malic enzyme. We also report the substrate activities of 3-hydroxyisocitrate and 3-fluoroisocitrate analogues. MATERIALS AND METHODS Enzymes. Type IV porcine heart NADP isocitrate dehydrogenase (Sigma) was used in all experiments. The lyophilized powder was dissolved in a minimum amount of 50 mM Pipes, pH 7.0, 50% glycerol, 10 mM EDTA, and 1 mM dithiothreitol and dialyzed against the same buffer for 48 h at 0 "C. The solution was then chromatographed on a 1 cm X 20 cm Sephadex G-10 column eluted with 10 mM Pipes, pH 7.0, 10% glycerol, and 1 mM dithiothreitol. The coupling enzyme glutamate dehydrogenase was also diluted in 50 mM Pipes, pH 7, and 50% glycerol and chromatographed as above. 13CKinetic Isotope Effects. 13Cisotope effects on NADP isocitrate dehydrogenase were determined by the method of O'Leary (1980). To provide adequate buffering capacity at low pH values, reactions below pH 6.0 were run in a total volume of 280 mL rather than the standard 20-mL assay volume. The following buffers were used at a final concentration of 150 mM: pH 4.0-5.5, potassium acetate; pH 5.5-6.5, Mes; pH 6.5-7.5, Pipes. Because of the acid denaturation of isocitrate dehydrogenase, aliquots of fresh enzyme had to be periodically added to the reaction vessel to achieve the desired fraction of reaction. The enzyme was allowed to exhaust the NADP present to stop the reaction at the desired point. Because the K, for NADP is less than 1 pM, the NADP concentration remained well-above saturation until the nucleotide was nearly exhausted. Hence, the external commitment was not significantly altered by this practice. Acid degradation of the NADPH produced and the low concentration of C 0 2 prevented reversal of the reaction. Above pH 6.0, glutamate dehydrogenase was used to convert the 2ketoglutarate produced to L-glutamate, as well as to recycle the nucleotide. The concentration of reagents was as follows: for complete conversion (20-mL volume), 8 mM three,erythro-DL-isocitrate (or isocitrate analogue), 0.4 mM NADP, 30 mM NH4Cl, and 2 mM MgC1,; for low conversion (with recycling of nucleotide at neutral pH, 20-mL volume), 100 mM threo,erythro-DL-isocitrate(or analogue), 1 mM NADP, 30 mM NH,Cl, and the specified MgC12 concentration; for low conversion (without recycling of nucleotide at neutral pH, 280-mL volume), 10 mM threo,erythro-DL-isocitrate(or analogue), 1 mM NADP, and the specified MgC12 concentration. A total of 100 units of glutamate dehydrogenase was used in those assays with the nucleotide recycle. Enough isocitrate dehydrogenase was added to obtain the desired fraction of reaction in 1-2 h. As much as 20 mL of buffer containing 3000 units of enzyme had to be added to those assays at the acid extreme of the pH profile. The vessels were maintained at 25 "C during turnover. For each assay with isocitrate or 3-fluoroisocitrate, the amount of COz produced (measured manometrically following isolation) corresponded to the expected extent of reaction. The reactions were quenched with 2-5 mL of concentrated H2S04. The 1zC/'3C
2936 B I O C H E M I S T R Y ratio of the C 0 2 was analyzed on a Nuclides Associates RMS 6-60 isotope ratio mass s p e a m m a equipped with a dual-inlet system or a Finnigan 6-Eisotope ratio mass spectrometer. Initial Velocity Studies. AI kinctic studies were performed by monitoring changes in NADPH concentration at 340 nm with a Beckman DU monochromator equipped with an Update Instruments OD converter. The cuvette compartment was maintained at 25 "C with thermospacers. Routine assay conditions were 100 mM buffer (same as specified above for different pH ranges plus Tridnc, pH 7.5-8.5),0.2 mM NADP, and the specified amount of MgC12. Because the K,,, for isocitrate is submicromolar at neutral pH, lO-cm path-length quartz cuvettes were used. When it was desired to maintain the Mg2+ at a specific low level, the following dissociation constants were employed: Mg-NADP = 19.1 mM (Apps, 1973); Mg-isocitrate = 6.64 mM. Synthesis of Substrates. (A) threo,eryfhro-DL-Isoitrate2-d. Triethyl oxalosuccinate was synthesized according to the procedure of Friedman and Kosower (1955) in 97% yield. Deuteriated isocitrate was produced by reducing triethyl oxalosuccinate with NaBD, under the conditions of Limburg and O ' h r y (1977). No attempt was made to influence the stereochemical course of the reduction; hence, the final product contained all four isomers of isocitrate. The isocitric lactone was hydrolyzed in 1 N KOH at 25 "C over 48 h and loaded onto a 3.5 cm X 30 cm Dowex AG 1-X2 cdumn in the formate form. Isocitrate was eluted with a 500-mL &2 N HCOOH gradient. 'H NMR showed 97% deuteriation at C2. Overall yield was 62%. To account for the prtsence of three inactive isomers of isocitrate, unlabeled isocitrate w u synthesized by the same procedure: IH NMR (270 MHz, D20) 6 2.00 (s), 2.5 (m), 4.05 (d). (B) threo,eryrhro-~~-3-Hydroxyisocitrate. 3-Hydroxyb citrate (1,2-dihydroxy-l,2,3-propanetricarboxylic acid, also called garcinia acid) was synthesized according to the procedure of Plaut et al. (1975) in which frons-aconitic acid is oxidized by osmium tetroxide in the presenced of potassium chlorate. The product was purified by the same chromatographic procedure described for isocitrate: 'H NMR (D20) 6 2.73 (d, 2 H), 4.00 (s, 1 H). Two isomers were present in the final preparation, but only one is enzymatically active. Deuteriated 3-hydroxyisocitrate was synthesized by the same procedure, but starting from perdeuterio-trans-aconitic acid. This was synthesized by incubating unlabeled trans-aconitic acid in D 2 0 at 80 "C for 60 h. This exchanged all hydrogen for deuterium by a 1,3-prototrophic shift. The product, 2,4,4perdeuteriogarcinate,was 99%dcuteriatcd by 'H NMR integration against an internal acetate standard. (C) 3-Flwroisocitrafe. 3-Fluoroisocitrate was synthesized according to the procedure of Gottwald and Kun (1965). The synthesis and characterization of the precursor, triethyl 3fluorooxalosuccinate, is also reported in Grissom and Cleland (1988). A total of 2.5 g of triethyl 3-flwrooxalosuccinate was dissolved in 5 mL of ethanol and added to 8.25 mL of H20. The pH was adjusted to 7 with KOH and cooled to 0 "C. A total of 0.206 g of NaBH, was dissolved in 8.25 mL of H 2 0 at 0 "C and added dropwise over 2 h. The pH was maintained at 7-8 with dilute H2S04. The workup from this point on is critical for not losing fluorine. The ether extract was dried over Na2S04and the solvent removed at 25 "C. 'H and 19F NMR showed this to be a complex mixture of at least five fluorine-containing compounds. The esters were hydrolyzed with 2:l acetic acid:concentrated HCI for 3 days at 25 "C. Progress of the hydrolysis was followed by TLC on silica gel with CHC13 and visualized by UV and Fe(N03)3/NH20H
GRISSOM A N D CLELAND
spray. The acid was removed by the addition of 2 volumes of H 2 0and rotary evaporated at 25 "C. This procedure was repeated until no acid could be detected by olfaction (ca. 6 times). The lactone of 3-fluoroisocitrate was hydrolyzed by incubation at pH 12.5 for 6 h at 25 "C (negligible F was lost during this time). The product was chromatographed on the same ion-exchange column described above, except the material was eluted with 2 L of 0-1 N KCI, yielding a peak with NADP isocitrate dehydrogenase activity. I9F NMR showed two groups of peaks with eight lines each at -163.7 ppm and -167.9 ppm relative to CFC13 at 0 ppm. These are the two diastereomeric pairs pwible with 3-fluoroisocitrate. Determination of Enzymatic Products: 3-Hydroxyisocitrate. ( A ) "C NMR. A 5-mL volume of 92 mM rhreo-r>3-hydroxyisocitrate (containing equal concentrations of the three other isomers of 1,2&ydroxy- 1,2,3-propnetricarboxylic acid) was incubated with 100 mM NADP, 20 mM MgCI2, and ca. 90 units of NADP isocitrate dehydrogenase at pH 7 and 25 "C. After 7 h, the solution was filtered through glass wool and transferred to a 12-mm NMR tube with a 5-mm coaxial D20 lock containing 1.4-dioxane as a standard (67.47 ppm). I3C NMR spectra at 50.3 MHz (on a Nicolet 200MHz multinuclear spectrometer) were acquired at 8 and 12 h after the addition of enzyme. A relaxation delay of 2 s was used to ensure complete relaxation of the carbonyl and carboxyl carbons following a 90" pulse. A spectrum of the reaction components but without isocitrate dehydrogenase served as a control. (B) NaBH, Trapping Experiments. The products of enzymatic oxidation of 3-hydroxyisocitrate in various stages of decarboxylation were trapped by reduction of the resulting carbonyl compounds with tritium-labeled NaBH, or NaBD,. The resulting negatively charged s p i e s were isolated by anion-exchange chromatography, and their structure was determined by 'H and 13CNMR. The following were contained in a 200-mL reaction volume at pH 7: 100 pmol of rhreo-~-3-hydroxyisocitrate (and equal concentrations of the other three isomen of 1,2&ydroxy- 1,2,3-propanetricarboxylic acid), 100 pmol of NADP, 10 mM MgC12, and 50 units of NADP isocitrate dehydrogenase. The progress of the reaction was monitored by the appearance of NADPH at 340 nm. When ca. 65% of the rhreo-bgarcinate had been converted to products (40min), the reaction was quenched by cooling to 0 "C; 3.0 g of Bio-Rad Chelex 100 was added to further inhibit the enzymatic reaction. The enzyme was removed by ultrafiltration with an Amicon UM-1 membrane. The resulting solution was divided into two 100-mL portions for reduction by 200 mg of NaBH, or NaBD, (each trace labeled with 10 mCi of )H) dissolved in 15 mL of H 2 0 at 0 "C. The borohydride solution was added dropwise over 20"in at 0 "C, with the pH maintained between 7.5 and 9 by the addition of dilute HfiO,. The solutions were further stirred for 1.5 h to ensure reaction of all materials. Several drops of formic acid were added to decompose residual borohydride. At pH 9, the solutions were loaded onto separate columns of Bio-Rad AG-MP-1 ( 1 .5-cm diameter, total bed volume = 42 mL) in the formate form. The columns were washed with 150 mL of H20, and the bound material was eluted with a 500-mL 0-2 N formic acid gradient. Isotope Effect Nomenclature. The nomenclature of Northrop (1977) is used in which isotope effects on a kinetic or thermodynamic parameter are indicated by a leading superscript (D, T, or 13 for deuterium, tritium, or I3C). Thus,"Kq is K,(H)/K,(D), while 1 3 & is the ratio of rate constants for I2C- and I3C-containing substrates.
VOL.
N A D P ISOCITRATE DEHYDROGENASE ISOTOPE EFFECTS
Data Analysis. Hyperbolic saturation curves were fitted to eq 4 with a BASIC version of Cleland's HYPER0 program2 (Cleland, 1977). 0 = VA/(K A) (4)
+
Tritium and I3C isotope effects determined by the internal competition method were computed with eq 5 . In eq 5 , RJ 1 3 ( V / K ) = 1% (1 -f)/log [1 -f(R//Ro)I (5) is the I3C/I2C ratio at fractional reaction f,and Ro is the I3C/I2C ratio of the starting material. pH profiles which exhibited a rate plateau at both low and high pH were fitted to eq 6 with the FORTRAN program (6) rate = log [[YL (YH)K/tJI/(l + K/H)]
+
WAVL, where YL and YH are the rates at low and high pH, respectively, K is the thermodynamic dissociation constant for the titratable group, and H is the hydrogen ion concentration. Data for deuterium isotope effects were fitted to eq 7 which rate = V A / [ K (1 + F,Ev/K) + A( 1 + F,Ev)] (7) assumes isotope effects on V and V / K , where F, is the fraction of deuterium label in the substrate, while Evand E V j K are the isotope effects minus one for the respective parameters. Partitioning of Oxalosuccinate. The partitioning of the intermediate oxalosuccinate was determined in a manner analogous to that for oxalacetate (Grissom & Cleland, 1985). The production of isocitrate and 2-ketoglutarate from oxalosuccinate was followed spectrophotometrically to determine the total directly the partitioning ratio RH (also equal to Cf, forward commitment for decarboxylation). Oxalosuccinate reduction to h i t r a t e was monitored by utilization of NADPH at 340 nm. Oxalosuccinate decarboxylation to 2-ketoglutarate was followed by the disappearance of oxalosuccinate at 281 nm (the isosbestic point for conversion of NADP to NADPH). Subtracting the rate of isocitrate production from the total rate of oxalosuccinate utilization gave the amount of 2-ketoglutarate produced. The apparent oxalosucciiate extinction coefficient at 281 nm was determined at both pH values. To prevent reversal of reaction 2, 2-ketoglutarate was coupled to the production of glutamate with glutamate-pyruvate transaminase. This equilibrium was shifted in favor of glutamate production by the presence of 100 mM L-alanine in the assay. The barium salt of oxalosuccinate was obtained from Sigma and converted to the sodium salt with Dowex AG 50-Na+. The contaminating 2-ketoglutarate was quantified with glutamate-pyruvate transaminase and lactate dehydrogenase in the presence of L-alanine and DPNH. This assay was found to give better results than glutamate dehydrogenase in the presence of oxalosuccinate. The oxalosuccinate concentration was determined by conversion of all oxalosuccinate to 2ketoglutarate by heating to 90 "C for 15 min. The instrumentation and methods used are otherwise the same as described in Grissom and Cleland (1985). The equation for calculating the intrinsic 13Cisotope effect from the partitioning ratio and the observed "C isotope effect is 13k = R H [ " ( V / K )- 11 + " ( V / K ) (8) where RH is the 2-KG/Ic partitioning ratio. RESULTS 3-Hydroxyisocitrate Product Identification: I3CN M R . The progress of the enzymatic oxidation was followed by the appearance of peaks at 23 and 173 ppm (corresponding to the Compiled FORTRAN versions of these programs for the IBM PC are available from C.B.G.
27,
2937
NO. 8, 1988
Table 1: "C NMR Resonances Ascribable to Decarboxylation Products of 3-Hydroxyisocitrate 8h 12 h structure chem shift (ppm) intensity chem shift (ppm) intensity 6 49.38 1 .o 49.37 1.3 49.66 64.87 68.50 173.95 175.06 175.23 206.84 209.40
4
4 6 4 6 4
4 6
1.7 1.2 0.9 0.6 0.6 0.4 0.5 0.6
49.66 64.86 68.49 173.95 175.06
0.4 1.o I .2 0.5 0.6
209.39
0.6
Relative intensities of resonances.
cool
n-c-on I -m-c-on
yI COOI
* NADPH
NADP
COO' I
c-on II c-on
co,
I y, COO-
1
COOI n k I- o n
n'-
c=o I CHI
z
I
n I n'-c-on I c =o I CHI I
-
COO-
COO' I
F=O
n'-c-on
FI -on
c=o I
?HI
7"I
coo-
coo-
I
-l n n*-Lon I c=o I
$"I
coo-
6
n. I c=o I cnon I
y 2 COO'
-1
FIGURE 1 : Scheme for 3-hydroxyisocitrate decarboxylation. I t is thought only the conversion of 1 to 2 is catalyzed by NADP isocitrate dehydrogenase. Subsequent steps are thought to occur in solution.
C, methylene and exocyclic amido carbons of NADPH), indicating the reduction of NADP. Resonances corresponding to the acid-catalyzed degradation product of the reduced nicotinamide ring were also observed in the 8-h spectrum and increased in intensity in the 12-h spectrum. These resonances were assigned according to the published values of Williams et al. (1976). Those resonan& not corresponding to NADP, NADPH, the acid-catalyzed degradation product of NADPH, unreacted 3-hydroxyhitrate, glycerol (added with enzyme), Pipes, or dioxane were tabulated and their intensities noted (Table I). Those resonances which decrease in going from the 8- to the 12-h spectrum are ascribed to an initial reaction product which decarboxylates to yield the resonances which increase in intensity. The resonances listed in Table I match those expected for structures 4 and 6 as detailed in Figure 1 . Two methylene resonances in magnetically similar environments are seen at 49.38 and 49.66 ppm. C3-C4-Cs of compound 4 is identical with C3-C4-Csof compound 6. Hence, the resonance of the methylene group in 4 and 6 should be nearly identical. This is in contrast to the methylene group on 3, which is not a to a carbonyl group and should be shifted upfield (cf. the, I3Cresonances for the methylene group in 6 which shifts upfield by 7 ppm upon reduction of the carbonyl in the reduction trapping compound, H-I). This suggests the resonances in Table I ascribed to 4 are indeed 2-hydroxy-3ketoglutarate and not 3 (2-kete3-hydroxyglutrate). Structure
2938 B I O C H E M I S T R Y
GRISSOM A N D CLELAND
4 is seen in the spectrum taken after 8 h and diminishes in the
12-h spectrum. The resonances representing structure 6 increase in intensity in the 12-h spectrum. This suggests the carbon skeleton proceeds from 3-hydroxyisocitrate to 4 and then to 6. No evidence for acetol (8), which is the logical decarboxylation product of 6, was observed. Acetol has 13C resonances at 25.9, 68.5, and 213.08 ppm under the above experimental conditions. Plaut et al. (1979, however, clearly demonstrated the presence of acetol (8) in enzymatic reaction mixtures of 3-hydroxyisocitrate by isolating the 2,4-dinitrophenylhydrazone derivative, methylglyoxal 2,4-dinitrophenylosazone. Reduction Trapping Experiments. Three separate peaks contained 3H were observed in the eluent. These peaks are labeled H-I and D-I for the earliest eluting peak from the NaBH4 and NaBD4 reductions, respectively. Similarly, the peaks which eluted second are H-I1 and D-11, and the third peaks are H-I11 and D-111. There was a fourth, less distinct radioactive peak, but this coincided with an increase in A,,, indicating nonspecific labeling of the nucleotide or its decomposition products. Each peak was rotary evaporated to dryness and redissolved in H20 to remove formic acid. This process was repeated until no HCOOH was detected by olfaction. Each peak was lyophilized from D 2 0 in preparation for 'H and I3C NMR. Peaks H-I1 and D-11, although containing some radioactivity, contained no organic material with an observable 'H or 13CNMR signal and are attributed to an inorganic contaminant of the tritiated sodium borohydride. 'H NMR in D 2 0 at 270 MHz of the other peaks showed the following: for H-I, 6 3.03 (dd, 1 H, JBc = 7.9 and JC4,= 14.9 Hz, CH2), 3.10 (dd, 1 H, JB4,= 5.5 and Jcct = 14.9 Hz, CH2), 4.20 (dd, 1 H, JA-B = 6.6 and JA-At = 11.7 Hz, CH20), 4.33 (dd, 1 H, JA,-B= 3.7 and JA-At = 11.7 Hz, CH20), 4.76 (m, 1 H, CHOH); for D-I, 6 3.01 (d, 1 H, Jcc = 14.9 Hz, CH2), 3.08 (d, 1 H, JcXt = 14.9 Hz, CHJ, 4.18 (d, 1 H, JA-At = 11.7 Hz, CH20), 4.31 (d, 1 H, JA-At = 11.7 Hz, CH20);for H-111, 6 2.21 (dd, 1 H, J = 3.3 and 15.2 Hz, CH2), 2.33 (dd, 1 H, J = 9.5 and 15.3 Hz, CH2), 2.36 (dd, 1 H, J = 5.8 and 13.1 Hz, CH2), 2.42 (dd, 1 H, J = 5.5 and 13.0 Hz, CH2), 4.2 (m, 2 H,CHOH),4.76 (m, 2 H,CHOHCOO); for D-111, 6 2.23 (d, 1 H, J = 15.1 Hz, CHJ, 2.35 (d, 1 H, J = 15.2 Hz, CH2), 2.37 (d, 1 H, J = 13.0, CH2), 2.44 (d, 1 H, J = 13.1), 4.68 (s, 1 H), 4.77 (s, 1 H). 13C NMR in D 2 0 at 50 MHz with 'H noise decoupling showed the following: for H-I, 6 42.04 (s, 1 C, CH3, 66.09 (s, 1 C, CH2O), 70.60 (s, 1 C, CHOH), 185.82 (s, 1 C, COOH); for H-111, 6 42.32 (s, 1 C, CH2), 71.28 (s, 1 C, CHOH), 75.48 (s, 1 C, CHOCOOH). Signal to noise ratio not sufficient to observe carboxyl carbons. Radioactive peaks H-I and D-I unambiguously correspond to the reduction product of 6. The 'H NMR of the lactone of this compound has previously been reported (Mori et al., 1979). No evidence for the reduction products of 5 or 7 (both of which should cochromatograph with 6 by ion exchange, but be distinguishable by NMR) was observed. Fractions H-I11 and D-I11 correspond to five-carbon species with two ionized carboxyl groups. The assignment of the structure of the reduction compound of 3 or 4 is not definite. Structure 6 (and subsequent acetol formation) could not arise from 2-keto-3hydroxyglutarate (3). Protonation of the enediol 2 at C3 prevents further decarboxylation of 3. Hence, the major product of enzymatic decarboxylation of 3-hydroxyisocitrate is 2-hydroxy-3-ketoglutarate (4). This results from protonation of the enediol, 2, at C2. In contrast, the enzyme protonates
'"t L
4
I
I
5
i
I
6
l
I
7
1
l
€3
PH
2: p& profile for oxalylglycine, a competitive inhibitor for isocitrate, versus pH. Assay conditions: 100 mM buffer, 0.1 mM dithiothreitol, 2.0 mM MgSO,, and 0.5 m M NADP. Data fitted to eq 6 . Calculated pK = 5.7 i 0.1, YL = 2.0 f 0.1, and YH = 8.8 0.3. FIGURE
*
the enolate resulting from isocitrate oxidative decarboxylation at Cj. Presumably, this enzymatic protonation is stereospecific. The stereochemistry cannot be determined from either H-I or H-111; however, since asymmetric induction during the reduction with borohydride is likely to be low, two diastereomers will result whether the enzymatic protonation at C2 was stereospecific or not. Isocitrate Analogue Product Identification. The stoichiometry of C 0 2 release to NADPH reduction was determined for 3-fluoroisocitrateand 3-hydroxyisocitrate to ensure monodecarboxylation. For 3-fluoroisocitrate, the C 0 2 / NADPH ratio (NADPH measured spectrophotometrically and C02measured manometrically following acid quench with H2S04)was 1.0, indicating its suitability for carbon isotope effect studies. The CO,/NADPH ratio for 3-hydroxyisocitrate was 2.8, indicating polydecarboxylation and its unsuitability for carbon isotope effect studies. The product of 3-fluoroisocitrateoxidative decarboxylation was shown to be 2-keto-3-fluoroglutarateby 19F NMR (Grissom & Cleland, 1988). The 19F NMR spectrum of threo,erythro-3-fluoroisocitrateshowed two groups of peaks with eight lines each at -163.7 ppm and -167.9 ppm (relative to CFC1, at 0 ppm). Upon incubation with isocitrate dehydrogenase, Mg2+, and NADP, the peaks at -167.9 ppm diminished in intensity, and a new set of eight lines was seen to appear at -190.2 ppm. This agrees with the chemical shift of authentic 2-keto-3-fluoroglutarate(Grissom & Cleland, 1988), and this assigns the upfield peaks as t h r e o - ~ ~ - 3 fluoroisocitrate. pKi Profile. Oxalylglycine, a competitive inhibitor of isocitrate (Northrop, 1974), was used to generate the pKi versus pH profile shown in Figure 2. The data were fitted to eq 6 with the following results: pK = 5.7 f 0.1, YL = 2.0 f 0.1, and YH = 8.8 f 0.3. This likely represents the true pK of the acid/base functional group on the enzyme which deprotonates the C2 O H to facilitate hydride transfer, although it could be another group whose protonation state is critical to catalysis. Because the rate is higher at neutral pH, the group must be ionized for catalysis. 13CIsotope Effects. The variation of the observed 13C isotope effect on V/KIc is shown numerically in Table I1 and graphically in Figure 3. The data were fitted to eq 6 with the following results: pK = 4.46 f 0.10, 13(V/KI,) plateau value at low pH is 1.040 f 0.004, and the 13(V/KIc)plateau at neutral pH is 1.0028 & 0.0005. The variation of the observed 13C isotope effect with deuteriated isocitrate is also shown in Table 11. When fitted to eq 6 (with the pK, fixed at 4.46 because the low-pH plateau is poorly defined by lack of data points), the results are as follows: I3(V/K&, plateau
NADP ISOCITRATE DEHYDROGENASE ISOTOPE EFFECTS
VOL. 27, N O . 8 , 1988
2939
Table 11: 13C IsotoDe Effects for NADP Isocitrate Dehydrogenase as a Function of pH with Mg2+ substrate' IC IC IC IC
PH 7.0 7.0 7.0 7.0
fraction of reaction (%) 13.0 7.3 12.8 9.4
high convb*c 11 810.3 11 810.3 11 810.3 11 810.3
low conp 11 768.4 11 784.0 11 767.7 11 789.3
I3(v/Kd 1.0038 1.0023 1.0039 1.0019 av 1.0030 f 0.0010 12.7 11 768.7 1.0038 IC 6.5 11 810.3 20.2 11 810.0 1.0038 IC 6.3 11 849.8 19.7 11 806.9 1.0040 IC 5.9 11 849.8 17.7 11 797.2 1.0049 IC 5.4 11 849.8 7.6 10777.8 1.0086 IC 5.3 10 866.7 7.6 10 772.8 1.0090 IC 5.3 10866.7 18.0 11 740.1 1.0104 IC 5.0 11 849.8 17.6 11 743.8 1.0100 IC 5.0 11 849.8 IC 4.85 11 849.8 15.0 11 724.6 1.0116 IC 4.6 11 849.8 17.8 11 622.3 1.0216 IC 4.3 11 849.8 17.1 11 595.3 1.0241 IC 4.3 11 849.8 17.3 11 573.2 1.0263 IC 4.1 11 849.8 4.3 11 537.8 1.0276 IC-2-d 7.0 11 899.0 10.0 11 894.0 1.0005 IC-2-d 7.0 11 899.0 10.0 11 893.0 1.0005 av 1.0005 f 0.0005 IC-2-d 6.33 11 899.0 17.0 11 890.0 1.0008 IC-2-d 5.82 11 899.0 6.7 11 887.0 1.0010 11 883.0 IC-2-d 5.45 11 899.0 6.3 1.0014 IC-2-d 5.34 10854.4 9.9 10811.1 1.0042 IC-2-d 5.28 11 899.0 8.1 11 850.0 1.0043 IC-2-d 5.03 11 899.0 5.4 11 849.0 1.0043 IC-2-d 4.98 11 899.0 6.1 11 850.0 1.0043 IC-2-d 4.71 11 899.0 4.7 11 838.0 1.0053 3-F-IC 7.0 11 022.7 13.0 10892.3 1.0129 3-F-IC 7.0 11 022.7 12.7 10893.0 1.0128 av 1.0129 f 0,0001 3-F-IC 5.2 11 022.7 8.4 10 829.9 1.0186 3-F-IC 5.2 11 022.7 8.1 10830.7 1.0185 av 1.0186 f 0.0001 The high-conversion isocitrate samples have a Substrate abbreviations: IC, unlabeled isocitrate; Ic-2-d, isocitrate-2-4 3-F-Ic, 3-fluoroisocitrate. the following standard deviations associated with them: 11 810.3 f 0.8, three determinations; 11 849.8 f 1.0, four determinations; 10 866.7 f 2.0, two determinations; 11 899 f 0.7, three determinations. cIsotopic ratios of 13C/12Chave been multiplied by lo6 and corrected for I7O content according to the geochemical ratio of 170/180 by subtracting 740 from the observed isotopic ratio.
1.04
1
Table 111: Viscosity Dependence of Kinetic Constants for Isocitrate Substrate Analogues, pH 7.0' qb K , (pM) Vn" VIK 1.0 3.8 f 0.7 0.77 f 1.01 0.20 f 0.03 2.9 1.4 f 0.4 0.090 f 0.004 0.06 f 0.02 3-F-IC 1.0 5.1 f 1.1 0.026 f 0.006 0.005 f 0.001 2.9 4.3 f 1.1 0.012 f 0.001 0.0028 f 0.0006 3-OH-IC 1.0 2.4 f 0.5 0.056 f 0.002 0.023 f 0.004 2.9 2.6 1.0 0.033 f 0.003 0.013 f 0.005 'Assay conditions: 0.5 mM NADP, 4 mM MgC12, and 100 mM Pipes, pH 7.0. Cuvettes with 10-cm path-length were used for all assays. bThe viscosity was increased by the presence of 30% sucrose to give an increase in relative viscosity of auuroximately 2.9.
substrate IC
-5
n >
1.02
.
1.01
.
I
*
n 4
I 5
6
7
PH
I3Cisotope effect on V/K for isocitrate versus H. Data fitted to eq 6 with the following results: pK = 4.46 f 0.10, r3(V/KIE) plateau a t low p H = 1.040 0.004, and the "(V/KIC) plateau at neutral p H = 1.0028 f 0.0005. FIGURE3: Observed
*
value at low pH = 1.016 f 0.002 and ' 3 ( V / K I cplateau )~ at neutral pH = 1.0009 f 0.0005.The 13Cisotope effect on V/K with 3-fluoroisocitrate is also presented in Table 11. 13( V / K3-F.Ic) at neutral pH is significantly above 1 and increases slightly at pH 5.2. This is the first 13Cor deterium isotope effect significantly greater than 1 with isocitrate dehydrogenase at neutral pH. Viscosity Dependence of Kinetic Parameters for Isocitrate Analogues. The kinetic parameters for isocitrate and its two
Table IV: Deuterium Isotope Effects for Isocitrate Analogues at pH 7' substrate DV ( v/Klc) IC 1.Ob 1.OOb 3-F-IC 1.o 1.45 f 0.15 3-OH-IC 1.o 1.53 f 0.13 'Assay conditions: 0.2 mM NADP, 2 mM MgC1, and 100 mM Pipes, pH 7.0. Cuvettes with 10-cm path-length were used for all assays. bData of Cook and Cleland (1981b), but verified in the present studv. ~
analogues were determined in the presence and absence of 30% sucrose as viscosogenic agent. The results are presented in Table 111. The velocities varied hyperbolically with substrate concentration in each case and were fitted to eq 4. The standard errors associated with some values are large because of the low K, values which approach the limits which can be
2940 B I O C H E M IS T R Y
GRISSOM A N D C L E L A N D
Table V: Partitioning of Oxalosuccinate at 25 OC" initial rate (uMlmin) 2-KG 2-KG/Ic 13( V/KdC 13kd 5.40 91 1.12 f 0.06 8.2 & 0.4 7.3 & 0.5 1.0065 & 0.0010 1.054 k 0.008 5.00 113 0.325 f 0.016 1.04 f 0.05 3.2 f 0.2 1.0109 f 0.0010 1.046 f 0.006 5.60 (Mn2+) 9.3 f 0.P 1.0094 1.096 f 0.014 "Assay conditions: 100 mM buffer, 0.3 mM NADPH, 30 mM L-alanine, 2 mM MgCI2, 5 mM OSA, and 40 units of glutamate-pyruvate transaminase in 0.5-cm cuvette. Averaged results of four determinations at each pH. bApparent extinction coefficient of oxalosuccinate (M-I cm-') under stated assay conditions. Cobservedcarbon isotope effect interpolated from data in Table I. dCalculatedintrinsic carbon isotope effect with the stated 2-KG/Ic partitioning ratio and I3((v/KI,) according to Grissom and Cleland (1985). c2-KG/Ic ratio calculated from Siebert et al. (1957) with 0.67 mM MnC1,.
PH
COSAb
IC
determined with 10-cm path-length cells. Deuterium Isotope Effects. The deuterium isotope effect was determined for isocitrate and its two analogues. The results are presented in Table IV. The data were fitted to eq 7, which assumes a deuterium isotope effect on Vand V/K. The determined isotope effect on Vis meaningless unless both deuteriated and unlabeled substrate preparations are free of inhibitors or contain equal concentrations of inhibitors. This condition may not have been met since three inactive isomers were present in each preparation. Hence, any isotope effect on V was ignored. Deuterium isotope effects on V/K are not affected by inhibitors. At pH 4.98 and 15 pM total MgZC (l/&J, D(V/KIc)= 1.06 f 0.03. This is not significantly different from the value of 1.OO observed at neutral pH and saturating Mgz+ by Cook and Cleland (198 1b). Partitioning of Oxalosuccinate. The method of Grissom and Cleland (1985) was used to quantitate the partitioning of oxalosuccinate between decarboxylation to 2-ketoglutarate and reduction to isocitrate. The averaged results of four determinations at pH 5.00 and 5.40 are presented in Table V. The 2-ketoglutarate/isocitrate(2-KG/Ic) ratio can be used with the observed 13(V/&) value at the same pH to calculate a value for the intrinsic 13C isotope effect (eq 8). A value abstracted from the literature (Siebert et al., 1957) for the 2-KG/Ic ratio at pH 5.6 and 0.67 mM Mn2+is also presented. The values of 13( V/K) used for the calculation of the intrinsic isotope effect at each pH are taken from the calculated parameters of eq 6. Hence, the standard errors associated with those values reflect errors distributed throughout the data in Table 11. Most of the error in the determination of 13kcomes from the error associated with the measurement of small l3(V/KIC)values. Below pH 5.0, the isotope effect does become larger, but the stability of NADPH becomes a problem such that considerable nucleotide is degraded by acid hydrolysis, thus interfering with the accurate determination of the partitioning ratio, R H . Above pH 6, the 2-KG/Ic partitioning ratio increases drastically, such that it becomes difficult to measure simultaneously decarboxylation and reverse hydride transfer rates, which differ by more than a factor of 15. DISCUSSION 3-Hydroxyisocitrate Product Identification: I3CNMR. The initial product of enzymatic oxidation and decarboxylation of 3-hydroxyisocitrate is a cis-enediol. By analogy to the accepted enzymatic mechanism with isocitrate, an enzymatic base protonates the enol to yield the ketone, which is subsequently released into solution. Removal of carboxyls C1 and C5 can only result from protonation at carbon 2 of the cisenediol, 2, to give the @-ketocarboxylic acid, 4. If carbon 3 of the cis-enediol were protonated, the resulting 2-keto-3hydroxyglutarate would be unable to decarboxylate further. Because isocitrate dehydrogenase can protonate either C2or C3, the base must be located between C2 and C3. The unexpected preference for protonation of the cis-enediol at C2
probably reflects the favorable thermodynamics of this molecule over 2-keto-3-hydroxyglutarate(i-e., the cis-enediol is electron-rich at C2, thus favoring protonation). Rose (1960) demonstrated the ability of isocitrate dehydrogenase to detritiate 2-ketoglutarate labeled with 3H at C3. This indicates a base on the enzyme does exist to protonate the enolate resulting from decarboxylationand the enzyme does not release the enolate into solution unprotonated. Carbon Isotope Effects. Figure 3 shows that the 13Cisotope effect increases significantly as the pH is lowered. This is the first report of an observed 13Cisotope effect greater than unity with Mg2+for NADP isocitrate dehydrogenase. Seltzer et al. (1959) first predicted the absence of an observed 13Cisotope effect in those reactions where the rate of catalysis exceeded the rate of substrate dissociation from the enzymesubstrate central complex. Indeed, this is thought to be the case with isocitrate dehydrogenase. Only if catalysis can be slowed relative to substrate dissociation would the observed isotope effect be increased as the commitment to catalysis is lowered. This can be accomplished by lowering the pH of the assay such that a smaller fraction of the enzyme acid-base catalyst responsible for deprotonating the C2 hydroxyl of isocitrate prior to hydride transfer will be in the correct protonation state for the reaction to occur. Cook and Cleland (1981a,b) fully developed the theory of the pH dependence of isotope effects. Consider the scheme EH
IrlA\
EAH
k0
1e
E
k,A k@Hllxe k
EA
A products
k2
where E is enzyme-NADP-Mg2+ and A is isocitrate. If k3 is greater than k2, then isocitrate will be sticky (i.e., dissociate from the central complex much less often than it goes on to form products). This is the case with NADP isocitrate dehydrogenase at neutral pH. At low pH, a large fraction of the enzyme population will be incorrectly protonated for catalysis to occur and deplete the E and EA pools, thereby increasing the tendency of isocitrate to dissociate from the central complex relative to catalysis. The observed pK for this titration will be displaced outward (toward the acid/base extremes) from its true value. The pKi profie for a competitive inhibitor of isocitrate will show this true pK, however, because there is no catalytic event to cause it to be sticky. Taking the true pK of 5.7 from Figure 1 and the stickiness-displacedpK of 4.46 from Figure 3, we calculate a displacement of 1.2 pH units. From eq 9, we can define a precise stickiness factor for P&pp PKtrue - 1% (1 -f- k 3 / k 2 ) (9) isocitrate. Hence, catalysis is 16 times faster than dissociation of isocitrate from the central complex. It is the internal competition nature of 13( V / K ) isotope effects which allows a pK as low as 4.46 to be precisely defined. Acid lability of the
VOL. 27, N O . 8 , 1988
NADP ISOCITRATE DEHYDROGENASE ISOTOPE EFFECTS
protein makes kinetics difficult even at pH 5.0. The initial velocity data of Cook and Cleland (1981b) at pH 4.5 suggested the presence of a pK about 4.5, but it is very difficult to determine such a low pK with initial velocity studies. The break in the V/K rate-pH profile seen at 6.73 (Cook & Cleland, 1981a,b) was attributed to the third pK of iswitrate. Hence, only triply ionized iswitrate is the catalytically active species. The plateau at low pH in Figure 3 represents the limiting value for the observed l3(V/KIC)for isocitrate after the external commitment has been removed. The limiting value of 1.040 for unlabeled isocitrate is about what would be expected for a decarboxylase with moderate commitments [cf. NADP malic enzyme 13( V/Kmal) = 1.03361. Presumably, only internal commitments to catalysis remain to diminish the intrinsic 13C isotope effect from its observed value. Since the observed value for l3(V/KIC)"is less than 13(V/KIc)H in all cases, the catalytic mechanism must be stepwise according to the multiple isotope effect theory of Hermes et al. (1982). The presence of deuterium increases the total commitment for decarboxylation, thus decreasing the observed 13(V/K). The multiple isotope effect method also allows the order of the deuterium- and "C-sensitive steps to be determined when "( V / K ) is known along with 13( V/KIc) with unlabeled and deuteriated substrates. If the deuterium-sensitive step precedes the 13C-sensitivestep, then eq 10 will be true. If the 13C-sensitivestep precedes the
- 1 =-" ( V / K ) 13(V/K)D - 1 "Keg
13(V/K)H
(10)
deuterium-sensitive step, then eq 11 will hold. Using the
13(V/K) values at the low-pH plateaus, eq 10 predicts3 a "(V/K)of 2.98 and eq 11 predicts a " ( V / K )of 2.87. There is no appreciable "( V / K )under any condition of low pH, low metal ion concentration,EDTA, NADPH, or 1 mg/mL bovine serum albumin, however. At pH 4.5, "(V/KIC) is only 1.08 f 0.02 (Cook & Cleland, 1981b). We do not understand why the relationship between "(V/K), 13(V/QH,and 13(V/KlD appears to break down with isocitrate dehydrogenase. Besides the original application to NADP malic enzyme, it has also been successfully applied to jade plant (Crassula) NAD malic enzyme (Grissom et al., 1987), 6-phosphogluconate dehydrogenase (Rendina et al., 1984), and prephenate dehydrogenase (Hermes et al., 1984). Partitioning of Oxalosuccinate. The pH dependence of the ratio of oxalosuccinate decarboxylation to reduction is illustrated in Table V. As the pH decreases, the external commitment decreases, thereby increasing the amount of iswitrate produced relative to 2-ketoglutarate. Table V shows the average intrinsic 13Cisotope effect on decarboxylation is 1.050 f 0.005 with Mg2+as metal ion. This is equal to the intrinsic 13Cisotope effect for decarboxylation of malate (1.049) in the analogous malic enzyme reaction (see preceding paper). This is not surprising, since both enzymes are thought to generate an enzyme-bound metal-coordinated @-ketocarboxylic acid with the leaving carboxyl group out of the plane defined by the 8-ketone carbonyl, the metal ion, and the oxygens of carboxyl Cl. At the low-pH plateau, I3(V/QH= 1.040 and I3(V/QD = 1.016.
"Kq = 1.0027 (OLeary & Yapp, 1978) and DKcq= 1.17 (Cook et al., 1980).
2941
Plaut (Siebert et al., 1957) determined the ratio of oxalosuccinate decarboxylase activity to isocitrate dehydrogenase activity and the ratio of isocitrate dehydrogenase activity to oxalosuccinate reductase activity with Mn2+ 30 years ago. From these ratios, we can determine the ratio of oxalosuccinate decarboxylase to oxalosuccinate reductase (2-KG/Ic) activity to be 9.6. The 13k calculated with this partitioning ratio (determined with Mn2+) and a l3(V/KIC)value of 1.0094 (also determined at pH 5.6 with Mn2+)is 1.096. This is larger than the maximum theoretical 13kof 1.07 expected by comparison of ground-state energies of I2C- and 13C-reacting molecules. Internal and External Commitments. The 13kvalues of 1.05 and the pH dependence of 13(V/K) allow the dissection of the total forward commitment for decarboxylation into internal (partition ratios of enzyme intermediates not affected by reaction conditions) and external (partition ratios involving substrate dissociation) components. The observed isotope effect is defined by the intrinsic isotope effect and the total forward commitment as given in eq 12, when the reverse I3k + Cf '3(V/K) = 1 + Cf commitment for C 0 2 is negligible.4 At neutral pH, Cf is 16.8. This includes internal and external contributions. At low pH, 13( V / K ) reaches a plateau value of 1.040, and according to eq 12, a forward commitment to catalysis of 0.25 remains. All external components of the commitment have been eliminated at this plateau, so the commitment is purely internal. Subfrom Cf+,,,, at neutral tracting & = 0.25 (1.5% of Cf-total) pH, we find C,,, = 16.5 (98.5% of Cf.total) at neutral pH. At neutral pH, the majority of the commitment is external in contrast to NADP malic enzyme where the internal commitment is similar in magnitude (0.348) but represents 75% of the total forward commitment for decarboxylation. Deuterium Isotope Effects. The deuterium isotope effect on V / K never increases above 1.08, even at pH 4.5 (Cook & Cleland, 1981b). It was thought decreasing the metal concentration to a subsaturating level would decrease the external commitment for isocitrate so that a significant increase in the D( V/KIc) value could be observed. If metal-isocitrate complex was the true substrate for the enzyme, however, decreasing the metal concentration would have no effect on the observed isotope effect. If the enzyme must bind either free metal or free malate first, then an isotope effect will be enhanced at low metal concentrations because the rate of catalysis will be less than the rate of dissociation of either substrate from the central complex. If the enzyme binds metal-isocitrate, however, the observed isotope effect will not be enhanced at low metal concentrations. The D(V/KI,) at pH 4.98 and 15 pM total Mg2+is 1.06-not significantlydifferent from Cook and Cleland's (198 1b) value of 1.OS at pH 4.5. This result agrees with other laboratories's claims that metal-isocitrate is the true substrate for NADP isocitrate dehydrogenase (Colman The full expression for 13(V/Krc)includes terms for the reverse commitment for isocitrate and the 13Kq: l3k '3(V/K)=
+ C&i"+ Cf.,, + Crl3Kq + Cf.in+ Cf.,, + c,
1
The reverse commitment in this mechanism is for the first product released, CO,.Since the enzyme's affinity for C 0 2is very low, the reverse commitment is assumed to be zero. This implies CO, is able to leave the active site immediately after decarboxylation, thus preventing reversal of decarboxylation when the CO,concentration is low. This assumption effectively removes the C, and I3Kqterms from the expression.
2942
GRISSOM AND CLELAND
BIOCHEMISTRY
& Villafranca, 1972; Colman, 1972), but an alternative explanation is that the binding of metal ion and isocitrate is highly synergistic. Free isocitrate is the substrate for yeast NAD isocitrate dehydrogenase (Gabriel & Plaut, 1986). Alternate Substrates. 3-Hydroxyisocitrate (garcinate) and 3-fluoroisocitrate are the only alternate substrates known for the oxidative decarboxylation reaction of NADP isocitrate dehydrogenase. Both isocitrate analogues are substituted at C3. The general structure is COOH
I I HOOC-C-
H-C-OH
I H-C-H
X
I
COOH X=H.OH, F
According to the kinetic constants in Table 111, the K,,, values are nearly the same for all analogues. V,,, and V/Klcare slightly lower for 3-fluoroisocitrate and 3-hydroxyisocitrate. For each substrate, V / K decreases when the viscosity is increased. The greatest decrease in V / K should occur for the stickiest substrate. The trends are qualitative because of the large errors associated with some of the parameters. Significant deuterium isotope effects are seen for 3-fluoroisocitrate and 3-hydroxyisocitratein Table VI. This is the first report of D(V/Krc) greater than 1.0 at neutral pH. Similarly, 13(V / K I c )for 3-fluoroisocitrate at neutral pH is also significantly greater than unity. 13(VI&,) for 3-hydroxyisocitrate would probably be of similar magnitude if nonenzymatic decarboxylation of the 2-hydroxy-3-ketoglutarate product did not interfere with measurement of the isotopic ratio of the enzyme-generated C 0 2 product. NADP isocitrate dehydrogenase appears to tolerate substitution only at the C3 position. Homoisocitrate and other analogues have been shown not to be substrates for the enzyme (Plaut et al., 1975). Concluding Remarks. The usefulness of the pH variation of isotope effects to dissect internal and external commitments has been illustrated. NADP isocitrate dehydrogenase, which shows a nearly nonexistent I3C isotope effect on V/K at neutral pH, exhibits a 13Cisotope effect on V / K which is only moderately diminished by commitments at low pH. An examination of the I3C isotope effect with deuteriated and unlabeled isocitrate has determined the catalytic mechanism to be stepwise with hydride transfer likely preceding decarboxylation. This generates an enzyme-bound @-ketocarboxylic acid intermediate which undergoes facile decarboxylation. The breakdown of eq 10 and 11 to describe the order of hydride transfer and decarboxylation is unexpected. The decrease in 13( V/Klc)is larger than predicted by the size of D( V/KIc).One possible explanation for the discrepancy is a change in the kinetic mechanism between the conditions of an internal competition experiment to determine 13( V/Klc)(high levels of enzyme, high concentration of Mgz+and isocitrate) and a direct comparison experiment to determine D( V/Klc) (low levels of enzyme, low concentrations of isocitrate, and high levels of Mg2+). Kelly and Plaut (198 1b) showed that the kinetic mechanism changes from random to ordered sequential with NADP binding last at low enzyme concentrations or in the absence of NADPH or EDTA. The experimental protocol to determine D( V/Klc)precisely matches these conditions. If the kinetic mechanism were ordered with isocitrate binding first, then at saturating NADP the commitment for isocitrate would be infinite, and no isotope effect would be seen. This
is not supported by our experimental data, however. D( V/KIc) does not increase in the presence of EDTA or NADPH. Perhaps a change to an ordered sequential kinetic mechanism under the conditions used to determine D( V/KIc)is not overcome by the presence of NADPH or EDTA. It is prudent to refrain from further speculation without additional experimental data and supporting theory. REFERENCES Apps, D. K. (1973) Biochim. Biophys. Acta 320, 379. Cleland, W. W. (1977) Methods Enzymol. 63, 103. Colman, R. F. (1972) J . Biol. Chem. 247, 215. Colman, R. F., & Villafranca, J. J. (1972) J. Biol. Chem. 247, 209. Cook, P. F., & Cleland, W. W. (1981a) Biochemistry 20, 1790. Cook, P. F., & Cleland, W. W. (1981b) Biochemistry 20, 1797. Cook, P. F., Blanchard, J. S., & Cleland, W. W. (1980) Biochemistry 19, 4853-4858. Dalziel, K., & Londesborough, J. C. (1968) Biochem. J. 237, 1739. Friedman, L., & Kosower, E. (1955) Organic Synthesis, Collect. Vol. 111, p 510, Wiley, New York. Gabriel, J. L., & Plaut, G. W. E. (1986) Fed. Proc., Fed. Am. SOC.Exp. Biol. 45, 1655. Gottwald, L. K., & Kun, E. (1965) J. Org. Chem. 30, 877. Grissom, C. B., & Cleland, W. W. (1982) Fed. Proc., Fed. Am. SOC.Exp. Biol. 41, 1151. Grissom, C. B., & Cleland, W. W. (1984) Biochemistry 23, 3347. Grissom, C. B., & Cleland, W. W. (1985) Biochemistry 24, 944. Grissom, C. B., & Cleland, W. W. (1988) Biochim. Biophys. Acta 916, 437. Grissom, C. B., Willeford, K. O., & Wedding, R. T. (1987) Biochemistry 26, 2594-2596. Hermes, J. D., Roeske, C. A., O’Leary, M. H., & Cleland, W. W. (1982) Biochemistry 21, 5106. Hermes, J. D., Tipton, P. A., Fisher, M., OLeary, M. H., Morrison, J. F., & Cleland, W. W. (1984) Biochemistry 23, 6263. Kelly, J. H., & Plaut, G. W. E. (1981a) J. Biol. Chem. 256, 330. Kelly, J. H., & Plaut, G. W. E. (1981b) J. Biol. Chem. 256, 335. Kurganov, B. I. (1967) Mol. Biol. (Moscow) 1, 17. Lienhard, G. E., & Rose, I. A. (1964) Biochemistry 3, 185. Limburg, J. A., & OLeary, M. H. (1977) Biochemistry 16, 1129. Mori, K., Takigawa, T., & Matsuo, T. (1979) Tetrahedron 35, 933. Northrop, D. B. (1977) in Isotope Effects on Enzyme Catalyzed Reactions (Cleland, W. W., O’Leary, M. H., & Northrop, D. B., Eds.) University Park Press, Baltimore. Northrop, D. B., & Cleland, W. W. (1974) J. Biol. Chem. 249, 2928. O’Leary, M. H. (1971) Biochim. Biophys. Acta 235, 14. OLeary, M. H. (1980) Methods Enzymol. 64, 83. OLeary,M. H., & Yapp, C. J. (1978) Biochem. Biophys. Res. Commun. 80, 155. Plaut, G. W. E. (1963) Enzymes, 2nd Ed. 7 , 105-122. Plaut, G. W. E., Beech, R. L., & Aogaichi, T. (1975) Biochemistry 14, 2581. Rendina, A. R., Hermes, J. D., & Cleland, W. W. (1984) Biochemistry 23, 6257.
Biochemistry 1988, 27, 2943-2952 Rose, Z. B. (1960) J . Biol. Chem. 235, 928. Seltzer, S., Hamilton, G. A., & Westheimer, F. H. (1959) J . Am. Chem. SOC.81, 4018. Siebert, G., Carsiotis, M., & Plaut, G. W. E. (1957) J . Biol. Chem. 226. 977.
2943
Uhr, M. L., Thompson, V. W., & Cleland, W. W. (1974) J . Biol. Chem. 249, 2920. Vickery, H. B. (1962) J . Biol. Chem. 237, 1739. Williams, T. J., Ellis, P. D., Bryson, T. A., Fisher, R. R., & Dunlap, R. B. (1976) Arch. Biochem. Biophys. 176, 275.
Two-step Internalization of Ca2+from a Single E-PCa2 Species by the Ca2+-ATPaset Daniel Khananshvilit and William P. Jencks* Graduate Department of Biochemistry, Brandeis University, Waltham. Massachusetts 02254 Received July 27, 1987; Revised Manuscript Received November 23, 1987
ABSTRACT: Phosphorylation by ATP of E.*Ca2 (sarcoplasmic reticulum vesicles (SRV) with bound 45Ca2+) during 5-10 ms leads to the occlusion of 2 *Ca2+/EP,,, [quench by ethylene glycol bis(0-aminoethyl ether)-N,N,N’,N’-tetraaceticacid (EGTA) alone] in both “empty” (10 p M free Ca2+in)or “loaded” S R V (20-40 mM free Caz+in).The rate of CaZ+“internalization” from the occluded E-P-*Ca2 was measured by using an ADP EGTA quench; a *CaZ+ion that is not removed by this quench is defined as internalized. In the presence of 20-40 mM unlabeled CaZ+inside SRV, 1 *Ca2+/EPtotis internalized from 45Ca-labeled E-P-*Ca2 with a first-order rate constant of kl = 34 s-l. Empty SRV take up 2 *Ca2+/EP,,, with the same initial rate, but the overall rate constant is k&sd = 17 s-l. The apparent rate constant (kb = 17 s-l) for internalization of the second *CaZ+is inhibited by [Cali,, with 1.3 mM and a Hill coefficient of n = 1 . 1 . These data show that the two CaZ+ions are internalized sequentially, presumably from separate sequential sites in the channel. [3ZP]EP*Ca2 obtained by rapid mixing of E C a 2 with [y-32P]ATPand EGTA disappears in a biphasic time course with a lag corresponding to -34 s-’, followed by EP* decay with a rate constant of 17 s-l. This shows that both CaZ+ions must be internalized before the enzyme changes its specificity fof catalysis of phosphoryl transfer to water instead of to ADP. Increasing the concentration of ATP from 0.25 to 3 mM accelerates the rate of 45Ca2+internalization from 34 to 69 s-’ for the first CaZ+ and from 17 to 34 s-l for the second Ca2+. High [ATP] also accelerates both phases of [32P]EP-Ca2 disappearance by the same factor. The data are consistent with a single form of ADP-sensitive E-P.Ca2 that sequentially internalizes two ions. The intravesicular volume was estimated to be 2.0 pL/mg, so that one turnover of the enzyme gives 4 mM internal [CaZ+].
+
-
-
T e mechanism of CaZ+transport by the Ca2+-ATPase1of sarcoplasmicreticulum may be described by a set of rules that define vectorial and chemical specificity in the coupled process (Jencks, 1980, 1983; Pickart & Jencks, 1984). The rule for vectorial specificity is that phosphorylation of the Asp-COOof the Ca2+-ATPasechanges the direction of the Ca2+dissociation-binding to the enzyme (Verjovsky-Almeida et al., 1978; Dupont, 1980; Takisawa & Makinose, 1981, 1983), acting as a vectorial switch that tells the enzyme whether to react with calcium outside or inside (Scheme I). Before phosphorylation Ca2+dissociates outside but not inside (Ikemoto, 1975; Inesi et al., 1980), whereas after phosphorylation it dissociates inside from the “occluded’’ EP.Ca2 but not outside (Dupont, 1980; Takisawa & Makinose, 1981, 1983; Champeil, et al., 1986). The rule for chemical specificity is that dissociation-binding of CaZ+acts as a chemical “switch” that tells the enzyme whether to react with ATP or Pi: In the absence of CaZ+the enzyme can be phosphorylated by Pi but not by ATP, and ECa2 can be phosphorylated by ATP but not by Pi (Sumida et al., 1978; Scofano et al., 1979; de Meis, 1981; Martonosi & Beeler, 1985). After phosphorylation of ECa, by ATP,
Scheme I
-ATP
/
inside
1
HpO-PI
all of the EPCa2 is ADP sensitive but not H 2 0 sensitive (Pickart & Jencks, 1982; Stahl & Jencks, 1984, 1987; Fernandez-Belda & Inesi, 1986). After calcium comes off EP.Ca2, the covalently bound phosphate can be transferred Abbreviations: E or CaZ+-ATPase,calcium adenosinetriphosphatase;
SRV, sarcoplasmic reticulum vesicles; EGTA, ethylene glycol bis(8‘Publication No. 1645. This research was supported in part by grants from the National Institutes of Health (GM20888) and the National Science Foundation (PCM-8117816). *Supported by a Chaim Weizmann postdoctoral fellowship.
0006-2960/88/0427-2943$01.50/0
aminoethyl ether)-N,N,N’,N’-tetraaceticacid; MOPS, 4-morpholinepropanesulfonic acid; HEPES, 4-(2-hydroxyethyl)-1 -piperazineethanesulfonic acid; Tris, tris(hydroxymethy1)aminomethane; BSA, bovine serum albumin; TNP, trinitrophenyl.
0 1988 American Chemical Society
